Very long baseline interferometry (VLBI) measurements of the sizes of compact extragalactic radio sources, jetted active galactic nuclei, provide data for probing the angular size-redshift relation, offering a classical cosmological test complementary to other distance-redshift methods. Aiming to update and extend previous studies conducted in the 1990s, we analyse a significantly expanded and improved dataset to reassess the angular size-redshift relation and its potential for constraining cosmological model parameters, focusing on the matter density parameter, Omega_m_, in a flat Lambda cold dark matter Universe. This is the first major update of the compact-source angular size test in the past quarter of a century, using a dataset an order of magnitude larger than in previous studies. We performed a Markov chain Monte Carlo analysis on real data and on multiple mock catalogues with varying Gaussian noise levels (10%, 20%, 50%) to evaluate parameter constraints in the presence of observational scatter. In addition, we conducted a test with 100 randomized catalogues created by shuffling redshifts while preserving other observables to explore the statistical significance of the angular size-redshift dependence. We also explored how astrophysical parameters depend on fixed cosmological models with different Omega_m_ values. The randomization test showed that the posterior distributions from randomized data do not overlap with those from real observations, with significant deviations, confirming that the measured angular size-redshift relation is physically meaningful and not a chance alignment. The astrophysical model parameter that describes the redshift dependence of the source angular size exhibits strong sensitivity and degeneracy with Omega_m_. Simulated mock catalogues indicate that the method is able to constrain Omega_m_ if the data scatter is below ~20%, but current real data noise levels are too high for reaching competitive cosmological constraints. Scaling estimates suggest that high-quality data of samples of several thousands to ~100000 sources, a standardization calibration approach, and/or refining sample selection criteria are needed to fully exploit the potential of the angular size- redshift test with this type of object.